An early morning outing

Prompted by a piece in Wednesday's Guardian (reference 1), curious about what meaning a price change since 1209 might have (£10 then equals £16,427.96 now) and having once done time, more than thirty years ago now, very near the bit of the Department of Employment which then did the RPI, I did a little digging.

My first stop was the Bank of England website at reference 2, where I completely failed to find the widget concerned. In fact, I thought that the search facility provided there was pretty hopeless, at least as far as this widget was concerned. So I asked Bing, and he turned it up straight away at reference 3, deep in the monetary policy part of the same website.

I then wanted to know what was so terribly wrong with the RPI and tried the ONS website at reference 4. Where I got on much better, quickly finding my way to references 5 and 6.

One difficulty is that these price indices, particularly the RPI, are built into all kinds of important calculations and you cannot just change the way you do the sums because they have fallen out of statistical fashion or because you are having a slack Friday afternoon. You have to go through a whole mountain of due process.

A second difficulty is scope. What things are in the index and what things are out? Which people are buying the basket of goods & services that your are pricing? The outcome of which is a range of indices, each meeting some particular need. The trick being not to have too many of them because that becomes confusing for everybody.

A third difficulty arises from the incomplete information about prices which goes into indexes of this sort. At the lowest level, say the price of a 500g tin of low-sugar baked beans, you get prices from a number of shops, perhaps a dozen or more, every week, every month or whatever. You then need to aggregate those price changes to give you one figure for baked beans. But you don't know the sales volumes by shop and so you can't do weighted averages, which would be proper - you have to do unweighted averages. And it seems that the RPI makes a lot of use of something called the Carli formula, invented by an Italian statistician some hundreds of years ago. A formula which has now been discredited and better formulae are available. But a formula which those on the worker side of wage negotiations like because it delivers a maximum estimate for inflation. 

Notwithstanding which, the RPI has been struck of the list of approved statistics and publication will cease in the not too distant future...

References

Reference 1: Initial values: How CPI, RPI and CPIH measure up - Larry Elliott, Guardian - 2023.

Reference 2: https://www.bankofengland.co.uk/.

Reference 3: https://www.bankofengland.co.uk/monetary-policy/inflation/inflation-calculator.

Reference 4: https://www.ons.gov.uk/.

Reference 5: Consumer price inflation detailed briefing note - Andy King, ONS - 2020.

Reference 6: UK Consumer Price Statistics: A review - Paul Johnson - 2015. Lots of good stuff here - and the source of the snap above.